OpenAI said on Wednesday that it has already exceeded its objective to secure 10 gigawatts of artificial intelligence computing capacity within the United States, accomplishing the milestone several years before the original 2029 target date set under its Stargate initiative.
The achievement comes just over a year after Stargate was launched, and OpenAI highlighted that more than 3 gigawatts of capacity were added in the last 90 days alone. The company framed the rapid build-out as a necessary step to train more sophisticated AI models, to scale deployments and to satisfy growing demand from businesses, developers and governments.
Stargate is described by OpenAI as a $500 billion AI infrastructure program intended to create the computing backbone for advanced AI. The initiative covers the development of large-scale data centres, power supply arrangements and high-performance chips. OpenAI said the effort is being carried out with partners that include Oracle (NYSE:ORCL), NVIDIA (NASDAQ:NVDA) and Microsoft (NASDAQ:MSFT), as well as various utilities and local governments.
In addition to announcing the 10 gigawatt milestone, OpenAI said it is evaluating additional data centre sites across the country as it looks to expand capacity further.
At the same time, OpenAI faces scrutiny about the sustainability of heavy infrastructure spending. A recent Wall Street Journal report said the company missed internal targets for revenue and user growth, including falling short of its goal to reach one billion weekly users for ChatGPT, and noted those shortfalls have raised concerns about OpenAI's ability to sustain high levels of compute investment.
The company did not provide new financial figures or a detailed timeline for further expansion in its announcement. Instead, the statement emphasized the connection between growing compute power and the operational needs of training and deploying larger AI models.
As OpenAI evaluates additional sites and continues to increase capacity, the company’s approach links large-scale infrastructure development with demand from customers across public and private sectors, while questions about revenue and user metrics remain part of the public discussion.